Re: [tcpm] Rechartering TCPM for alternative congestion control algorithms

[Joe Touch <[e-mail]>]

On 1/27/2015 7:56 AM, Scharf, Michael (Michael) wrote:
> Hi Joe,
>
> The current TCPM charter very explicitly includes "incremental
> enhancements of TCP's standard congestion control". RFC 5681 is a TCPM
> document.

Yes, where incremental would include things like tweaking RTO or ACK counts.

> Are you suggesting that *all* documents dealing with TCP congestion
> control should be moved from TCPM to TSVWG, including e.g. potential
> future updates of RFC 5681?

Major changes that aren't specific to TCP, IMO yes, and yes, that would 
definitely IMO include updates to RFC5681.

The community affected and interested in such changes is far beyond that 
of just TCPM. IMO, this is better for TCMP, TSVWG, and the development 
of overarching mechanisms.

> I guess this would result in a major re-chartering both of TCPM and
> TSVWG. And it could be a slippery road. A lot of TCPM documents affect
> congestion control at least partly.

We already have that sort of slippery slope issue. The general rule, 
AFAICT, is "minor" stays in TCPM and major goes to TSVWG. Major always 
includes changes that affect other transports as much as they affect TCP.

(we could even say that adopting CUBIC should happen in TSVWG, but that 
later minor tweaks could happen in TCPM)

Joe

>> -----Original Message-----
>> From: Joe Touch [mailto:[e-mail]]
>> Sent: Tuesday, January 27, 2015 4:34 PM
>> To: Scharf, Michael (Michael); [e-mail]
>> Cc: [e-mail]
>> Subject: Re: [tcpm] Rechartering TCPM for alternative congestion
>> control algorithms
>>
>> Hi, all,
>>
>> I disagree; TCP congestion control algorithms are also deployed in
>> other
>> transports, and thus this seems more appropriate for TSVWG.
>>
>> Joe
>>
>> On 1/27/2015 2:21 AM, Scharf, Michael (Michael) wrote:
>>> Hi,
>>>
>>> This e-mails asks for community feedback on a suggested small addon
>> to the TCPM charter [1].
>>>
>>> In the last TCPM meeting [2] there was strong support for adopting a
>> document describing the CUBIC congestion control algorithm [3]. To the
>> chairs, it is not entirely obvious whether this document, or possibly
>> other similar documents, would indeed be in scope of the current TCPM
>> charter. Given the importance of the TCP congestion control, we prefer
>> a community consensus explicitly documented in the charter instead of
>> ambiguity.
>>>
>>> The current charter limits the scope of TCPM to "modest changes to
>> the protocol, algorithms, and interfaces". It allows "incremental
>> enhancements of TCP's standard congestion control" but explicitly
>> mandates rechartering for fundamental changes [1]:
>>>
>>> OLD:
>>>
>>> TCPM also provides a venue for standardization of incremental
>>> enhancements of TCP's standard congestion control, but such changes
>>> may require additional review by the IRTF Congestion Control
>>> Research Group (ICCRG). Fundamental changes to TCP or its congestion
>>> control algorithms (e.g., departure from loss-based congestion
>>> control) will be handled by other working groups or will require
>>> rechartering.
>>>
>>> We suggest to update this paragraph in the TCPM charter by an
>> explicit statement that "TCPM may document alternative congestion
>> control algorithms that are known to be widely deployed, and that are
>> considered safe for large-scale deployment in the Internet":
>>>
>>> NEW:
>>>
>>> TCPM also provides a venue for standardization of incremental
>>> enhancements of TCP's standard congestion control. In addition,
>>> TCPM may document alternative congestion control algorithms
>>> that are known to be widely deployed, and that are considered
>>> safe for large-scale deployment in the Internet. Changes of
>> algorithms
>>> may require additional review by the IRTF Congestion Control
>>> Research Group (ICCRG). Fundamental changes to TCP or its congestion
>>> control algorithms (e.g., departure from loss-based congestion
>>> control) will be handled by other working groups or will require
>>> rechartering.
>>>
>>> In our reading, "TCP's standard congestion control" is currently
>> defined by RFC 5681.
>>>
>>> This e-mail and the suggested rechartering does not imply any
>> adoption of one or more alternative congestion control algorithms.
>>>
>>> Any feedback regarding this suggested rechartering would be very
>> welcome. In particular, please let us know if there are any concerns
>> with this proposal or if you have suggestions for a different wording.
>> Please let us know any thoughts until Feb. 15, 2015.
>>>
>>> Thanks a lot!
>>>
>>> Michael, Pasi, Yoshifumi
>>>
>>>
>>> [1] https://datatracker.ietf.org/wg/tcpm/charter/
>>>
>>> [2] http://www.ietf.org/proceedings/91/minutes/minutes-91-tcpm
>>>
>>> [3] https://datatracker.ietf.org/doc/draft-zimmermann-tcpm-cubic/
>>>
>>> _______________________________________________
>>> tcpm mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tcpm
>>>